In general, on the circuit board such as a PCB, various types of elements such as an integrated circuit are disposed and signal lines are provided between these elements so that a signal is interchanged through the signal lines between the elements. In this case, the level (electric potential) of the signal interchanged through the signal line between the elements takes an electric potential state which is one of a low potential state (Low state) and a high potential state (High state) to achieve the interchange of binary information of “0” and “1” between the aforesaid elements.
On the reception side (for example, an integrated circuit) of such a signal, a state discrimination is made as to whether the level of the signal propagated through the signal line is in a low potential state (Low state, i.e., “0”) or in a high potential state (High state, i.e., “1”). This state discrimination is made regarding a threshold (threshold electric potential) as the reference. For example, in a case where a common TTL (Transistor-Transistor Logic) level is employed as a signal level, the amplitude voltage of the signal is 5V, and the signal level (electric potential) of 0.8 V or less is judged to be in a low potential state, the signal level of 2.4 V or more is judged to be in a high potential state, and the signal level between 0.8 V and 2.4 V both inclusive is judged to be in a gray zone, which belongs to neither of these states. That is, the common TTL makes the discrimination on the low potential state and the high potential state, regarding the threshold (0.8V) for the low potential state discrimination and the threshold (2.4V) for the high potential state discrimination as references.
Meanwhile, as a technique for the state judgment using thresholds, for example, there has been proposed the technique disclosed in the following Patent Document 1. According to the disclosure of this Patent Document 1 calculates the similarity between an inputted voice pattern and each of a plurality of standard patterns and compares the calculated similarity with a predetermined threshold to accomplish a voice recognition. In particular, this Patent Document 1 discloses that the aforesaid threshold is strictly set (set to be a small value) in advance in order to avoid the erroneous recognition while the threshold is set to be a large value at the re-input of the inputted speech rejected on the basis of the initial threshold. That is, the threshold is dynamically set in accordance with the result of the voice recognition (reject situation).
Patent Document 1: Japanese Patent Laid-Open (KOKAI) No. HEI 4-152397